Five experiments were conducted in the summer of season 2002 to evaluate ten newly mungbean genotypes compared with the commercial cultivar Kawmy-1.  The experiments represent a wide range of environmental conditions to assess the best genotypes to replace the commercial cultivar.  The experiments were sown at Maryout in the middle of March and July; Shalakan in the middle of March and Sohag in the middle of March and July.  Data of number of seeds/plant, 100 seed weight and seed yield/plant were subjected to stability analysis proposed by Eberhart and Russell (1966) and Tai (1971).  The eleven genotypes showed a good performance at Shalakan and Sohag.  The mean values indicated that number of seeds/plant was the major controbutor to seed yield rather than seed weight.  The commercial cultivar Kawmy-1 was the best in No. of seeds/plant in all environments, however, it was the inferior in 100-seed weight.  Six genotypes were unstable in number of seeds/plant and 100-seed weight and showed significant deviation from linear response (S²d and ), however, their b's did not differ significantly from unity and their a were small.
                There was a lack of association between stability and high yielding ability.  The best two unstable genotypes which are likely to candidate to replace the commercial cultivar Kawmy-1 and significantly outyielded it by 34.55 and 40.30% are No. 6 (L3740) and No. 7 (L3940).  The only two stable genotypes; No. 5 (L3630) and No. 8 (L2020) outyielded the check by 13.68 and 20.40%, respectively.